Unexpected Parsoid tokenization
Closed, ResolvedPublic
Actions

Assigned To

Authored By

	Renxiaoyi
	Aug 17 2016, 2:55 AM

Description

From Parsoid page https://en.wikipedia.org/api/rest_v1/page/html/Dumb_and_Dumber, we see the following data-mw for the infobox:

data-mw='...,"gross":{"wt":"$247 million<ref name=\"mojo title"}},"i":0}}]}'

And the wikitext source for the "gross" attribute is

| gross = $247 million<ref name="mojo title|dumbanddumber">{{mojo title|dumbanddumber}}</ref>

It seems the Parsoid tokenizer incorrectly truncates the line at the first pipe char.

Mentioned In: T113044: Converting {{{foo}}} from wikitext to html to wikitext returns a 500 error
T102134: Content anchors (e.g. #cite_note5) points to Main_Page instead of current article
Mentioned Here: rGPAR974dd5b3d70a: T113044: Complete templatearg representation in spec
T102134: Content anchors (e.g. #cite_note5) points to Main_Page instead of current article
T113044: Converting {{{foo}}} from wikitext to html to wikitext returns a 500 error
T115186: Fix bug in reparseTableAttributes which is broken in some scenarios

Hmm, must be the precedence of extension tags, because we do the right thing if that were a div.

Related: T115186 ... I haven't looked closely if they are duplicates, but worth investigating.

Change 328444 had a related patch set uploaded (by Arlolra):
T143183: Extensions take precedence over templates

Change 328444 merged by jenkins-bot:
T143183: Extensions take precedence over templates

Arlolra closed this task as Resolved.Dec 21 2016, 7:20 PM